Evaluation of the reliability of pulse oximetry, at different attachment sites, to detect hypoxaemia in immobilized impala (Aepyceros melampus) ...
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Veterinary Anaesthesia and Analgesia 2020, 47, 323e333 https://doi.org/10.1016/j.vaa.2019.08.051 RESEARCH PAPER Evaluation of the reliability of pulse oximetry, at different attachment sites, to detect hypoxaemia in immobilized impala (Aepyceros melampus) Thembeka K Mtetwaa,b, Gareth E Zeilerc, Liesel Laubscherd,e, Silke Pfitzerf,g & Leith CR Meyera,b a Department of Paraclinical Sciences, Faculty of Veterinary Science, University of Pretoria, Onderstepoort, South Africa b Centre for Veterinary Wildlife Studies, Faculty of Veterinary Science, University of Pretoria, Onderstepoort, South Africa c Department of Companion Animal Clinical Studies, Faculty of Veterinary Science, University of Pretoria, Onderstepoort, South Africa d Wildlife Pharmaceuticals, Rocky Drift, White River, South Africa e Department of Animal Science, Faculty of Agriculture, University of Stellenbosch, Matieland, South Africa f Murdoch University, School of Veterinary and Life Sciences, Murdoch University, Perth, Western Australia, Australia g Wildlife Vets, Ngongoni Farm, Tipperary, South Africa Correspondence: Thembeka K Mtetwa, Department of Paraclinical Sciences, Faculty of Veterinary Science, University of Pretoria, P. Bag X04, Onderstepoort, Gauteng, 0110, South Africa. E-mail: thembeka.mtetwa@up.ac.za Abstract precision ¼ 3, Arms ¼ 4). The ear, cheek and rectal probes failed to give accurate or precise readings (ear: bias ¼ 4, Objective Evaluation of the reliability of pulse oximetry at precision ¼ 14, Arms ¼ 15; cheek: bias ¼ 12, precision ¼ four different attachment sites compared to haemoglobin 11, Arms ¼ 16; and rectum: bias ¼ 5, precision ¼ 12, oxygen saturation measured by a co-oximeter and calcu- Arms ¼ 13). lated by a blood gas analyser in immobilized impala. Conclusions and clinical relevance In order to obtain ac- Study design Randomized crossover study. curate and precise pulse oximetry readings in immobilized Animals A total of 16 female impala. impala, probes must be placed under the tail and SaO2 must be above 90%. Since SaO2 values are usually low in Methods Impala were immobilized with etorphine or thia- immobilized impala, pulse oximeter readings should be fentanil alone, or etorphine in combination with a novel interpreted with caution. drug. Once immobilized, arterial blood samples were collected at 5 minute intervals for 30 minutes. Then oxygen was insufflated (5 L minute1) intranasally at 40 minutes Keywords immobilization, impala, oxygen saturation, and additional samples were collected. A blood gas analyser pulse oximetry. was used to measure the arterial partial pressure of oxygen and calculate the oxygen haemoglobin saturation (cSaO2); a Introduction co-oximeter was used to measure the oxygen haemoglobin saturation (SaO2) in arterial blood. Pulse oximeter probes Chemical immobilization is an essential tool for wildlife con- were attached: under the tail, to the pinna (ear) and buccal servation and management worldwide. It is commonly used in mucosa (cheek) and inside the rectum. Pulse oximeter impala, mainly, but not exclusively, for translocation and dis- readings [peripheral oxygen haemoglobin saturation (SpO2) ease surveillance purposes. To immobilize impala, potent opi- and pulse quality] were recorded at each site and compared oids (such as etorphine or thiafentanil) are primarily used. with SaO2 and cSaO2 using Bland-Altman and accuracy of Impala are particularly sensitive to the adverse effects of these the area root mean squares (Arms) methods to determine the drugs and deaths often occur when they are captured (Meyer efficacy. P value < 0.05 was considered significant. et al. 2008; Zeiler & Meyer 2017a). These opioids impair ventilatory function by causing ventilatory depression, intra- Results Pulse quality was ‘good’ at each attachment site. pulmonary shunting, ventilation-perfusion mismatch and SpO2 measured under the tail was accurate and precise but probable impedance to oxygen diffusion, ultimately leading to only when SaO2 values were above 90% (bias ¼ 3, 323
Reliability of pulse oximetry in impala TK Mtetwa et al. severe hypoxaemia (Zeiler & Meyer 2017a,b). Untreated Materials and methods hypoxaemia may cause tissue hypoxia, organ failure and death (Fahlman 2014). Therefore, it is important that blood Animals oxygenation is continuously and accurately monitored during The study was approved by the Animal Ethics Committee of the immobilization to ensure that hypoxaemia is detected and University of Pretoria (V035-17) and was conducted at treated early to guarantee the welfare and survival of an Ngongoni Farm (25 31’25.2”S, 31 06’50.8”E). A total of 16 animal. free-ranging adult female impala (Aepyceros melampus), Currently, co-oximeters and blood gas machines are temporarily housed together in a 6 8 m outdoor enclosure frequently used for the detection of low arterial oxygen hae- (boma), were used. The impala [body weight 34.1 ± 5.2 kg d moglobin saturation (SaO2) or oxygen partial pressures or both mean ± standard deviation (SD)] were habituated to the (Wong et al. 2011; Bardell et al. 2017). However, these devices housing conditions 14 days prior to the study. Lucerne (Med- do not give continuous readings, they can be costly and the icago sativa) and water were supplied ad libitum with game arterial samples they require can be difficult to obtain in some pellets supplemented as needed. For exclusion, animals were species, including immobilized impala (Proulx 1999). checked for overt signs of cardio-respiratory and other diseases Furthermore, these devices are not always practical to use in at the first immobilization. the field (Proulx 1999). Pulse oximetry (SpO2), which gives an indirect peripheral measure of SaO2 (Pedersen et al. 2009), has Immobilization become a popular method used for the early detection of hypoxaemia in human and veterinary medicine (van de Louw Data for this study were collected during an unrelated study et al. 2001; Young et al. 2002; Giguere et al. 2014; Haymerle (see Pfitzer et al. 2019) which aimed to assess the dose- et al. 2016). It is widely used in the hospital and field setting response effect of the serotonin R-enantiomer of 8-hydroxy- because pulse oximetry devices are cheap, portable, easy to 2-(di-n-propylamino) tetralin drug on opioid-induced respira- apply and use, and give continuous noninvasive measures. tory depression. This ‘novel drug study’ used a randomized Their use and effectiveness have been determined and well crossover study design, in which the impala were immobilized documented in humans (Lee et al. 2000; van de Louw et al. (one at a time) on six different occasions. In each trial the 2001) and domestic animals (Chaffin et al. 1996; Mathews impalas were darted with either etorphine alone [0.09 mg et al. 2003; Giguere et al. 2014; Grubb & Anderson 2017; kg1, Captivon, Wildlife Pharmaceuticals (Pty) Ltd., South Reiners et al. 2018). However, although they are commonly Africa], thiafentanil alone [0.09 mg kg1, Thianil, Wildlife used during the immobilization of wildlife (Martin-Jurado et al. Pharmaceuticals (Pty) Ltd.] or etorphine (0.09 mg kg1) in 2011; Haymerle et al. 2016), their accuracy and utility have combination with different doses of the novel drug, using a 1.5 not been adequately determined. Therefore, the validity of their mL dart (P-type Pneudarts with 19 mm long needle; DAN- use for this purpose is still questionable. INJECT, International S.A, South Africa) projected into a Studies in human and domesticated animals show that the muscle mass of the pelvic girdle by a carbon dioxide (CO2) gas accuracy of pulse oximetry is dependent upon probe placement powered rifle [230 kPa: X-Caliber dart gun; Wildlife Pharma- (Chaffin et al. 1996; Lee et al. 2000; van de Louw et al. 2001, ceuticals (Pty) Ltd.]. The dart rifle operator stood on an Mathews et al. 2003; Giguere et al. 2014). In order to obtain a elevated walkway above the boma wall and darted the impala pulse oximeter reading, veterinarians working with immobi- over a distance ranging from 5 to 12 m. lized wildlife place probes at different sites including the ear, lip, rectum, tongue, vulva and the tip of the penis (Proulx 1999). Experimental procedures However, whether the readings from these sites are accurate Following immobilization, the impala were blindfolded and and precise still needs to be elucidated. Although several carried to an under roof workstation and placed in sternal re- studies have previously compared SpO2 readings at different cumbency on a table. A peripheral artery (either arteriae attachment sites in domestic animals (Chaffin et al. 1996; auricularis posterior or arteriae digitales palmares communes) was Mathews et al. 2003; Giguere et al. 2014), no such studies aseptically cannulated (22 gauge; Jelco Smiths Medical, UK) for have been performed in immobilized wildlife. arterial blood sampling. A total of four pulse oximeters (Nonin Therefore, this study aimed to establish, in immobilized PalmSat 2500 A; Kyron Laboratories (Pty) Ltd., South Africa) impala, whether pulse oximetry is a reliable method for the were connected, using four probes [Kyron Laboratories (Pty) determination of SaO2. The study also aimed to determine Ltd.] at different locations on the impala, as follows (Fig. 1): which attachment site of the pulse oximeter probe provided the most accurate SpO2 measurement. We hypothesized that pulse 1) Taildreflectance probe (2000T Transflectance probe; Kyron oximetry can be used to accurately measure SaO2 in immo- Laboratories (Pty) Ltd.) placed and secured gently, using adhe- bilized impala under field conditions. sive tape (25 mm ElastoPlast Tape; BSN Medical (Pty) Ltd., South 324 © 2020 Association of Veterinary Anaesthetists and American College of Veterinary Anesthesia and Analgesia. Published by Elsevier Ltd. All rights reserved., 47, 323e333
Reliability of pulse oximetry in impala TK Mtetwa et al. Africa), on unpigmented skin on the ventral aspect of the tail sensor and 8000R-reflectance sensor), which are compatible base within 20e30 mm from the root of the tail (Fig. 1a & a1). with each other (https://www.nonin.com/wp-content/ 2) Rectumdreflectance probe (2000T Transflectance probe) uploads/2018/09/Sensor-Compatibility-Guide.pdf). The de- inserted approximately 10 mm into the rectum and secured in vice’s set averaging time for SpO2 readings is 10 seconds position by taping the connecting wire of the probe to the tail base using adhesive tape (Fig. 1b & b1). (https://www.nonin.com/wp-content/uploads/Operators- 3) Eardtransmission probe (2000SL Lingual probe; Kyron Labo- Manual-2500A-Vet.pdf). ratories (Pty) Ltd.) clipped onto a shaved unpigmented area on In all the trials, 1 mL of arterial blood was collected in the caudal margin of the pinna (Fig. 1c & c1). heparinized plastic syringes (B Braun, Germany) and stored on 4) Cheekdreflectance probe (2000T Transflectance probe) secured ice at 5, 10, 15, 20 and 30 minutes from the time the impala to a holding clip and positioned inside the mouth and placed became recumbent. At 40 minutes of immobilization, seven against the buccal mucosa approximately 30 mm from the impala were given etorphine alone or thiafentanil alone and commissure of the lip (Fig. 1d & d1). were supplemented with oxygen for 10 minutes using a high- These sites were chosen based on practical application of the pressure oxygen cylinder [Ecomed Medical (Pty) Ltd, South probes. The Nonin PalmSAT 2500 veterinary probes were used Africa]. The flowmeter was set at 5 L minute1, administered in the study (2000T and 2000SL) because they are similar to intranasally via a nasal tube to increase the animal’s oxygen the Nonin PalmSAT 2500 human probes (8000Q2-ear clip haemoglobin saturation. Thereafter, arterial samples were also Figure 1 Placement of pulse oximeter probe at four different attachment sites. The tail (a) probe was a transflectance probe (a1). It was placed on unpigmented skin at the base of the tail and secured with tape. The probe inserted into the rectum (b) was a transflectance probe (b1). It was placed 10 mm inside the rectum facing the mucosa next to the temperature probe and indicated by the black arrow. The lingual transmission probe (c1) was clipped to the skin of the shaved margin of the ear (c) of the impala. The probe attached to the cheek (d) was a transflectance probe. It was attached to a plastic clip with rubber bands (d1) and the plastic clip secured the probe adjacent to cheek the 30 mm from the commissure of the lips. White dotted arrows represent internal probes placed on the mucosa and the white solid arrows are probes that are placed externally on the skin. © 2020 Association of Veterinary Anaesthetists and American College of Veterinary Anesthesia and Analgesia. Published by Elsevier Ltd. All rights 325 reserved., 47, 323e333
Reliability of pulse oximetry in impala TK Mtetwa et al. collected at random intervals during oxygen supplementation. random errordby calculating the SD of the differences), and Peripheral oxygen haemoglobin saturation (SpO2), pulse rate limits of agreement (LOA) (calculated as the bias ± 1.96 SD) at (beats minute1) and pulse quality data detected by the pulse 95% confidence interval for the mean difference and LOA. A oximeter were recorded in triplicate, within 30 seconds of frequency distribution test and column statistics were per- collecting each arterial blood sample. Pulse quality (pulse formed to determine whether the mean differences between signal strength) was indicated by a flashing coloured light the two variables (methods) were normally distributed. emitting diode built into the pulse oximeter; green indicated The combined accuracy and precision of the compared good, amber intermediate and red a poor quality. The pulse variables were determined by using the accuracy of the area oximetry readings were recorded in the following order: 1) tail; root mean square (Arms) (Batchelder & Raley 2007; Equation 2) rectum; 3) cheek; and 4) ear. (1) in Appendix A). According to ISO 80601-2-61:2011 guidelines, pulse oximetry is regarded as accurate and precise Arterial blood analysis when the bias, precision and Arms are ±3%, 3% and ±4%, respectively (DIN EN, 2011). After withdrawal, arterial blood samples were immediately stored on ice and analyzed within 10 minutes after collection Data handling using: 1) an EPOC blood gas analysis system and self- calibrating BGEM3 test cards (Epocal Inc., ON, Canada); and The SpO2 data set was paired with the SaO2 and cSaO2 data 2) a daily calibrated Avoximeter 4000 co-oximeter with cu- sets to determine the total number of pairs available for anal- vettes [Surgical Innovations (Pty) Ltd., South Africa)]. The ysis. Initially, all of the SpO2 measurements (all-data) were blood gas analyser was used to measure the arterial partial used to calculate the mean of the triplicates for each data point pressure of oxygen (PaO2) to calculate the arterial oxygen and these values were used for analyses using the methods haemoglobin saturation (cSaO2) using a default equation described. Then, exclusion criteria were applied to the SpO2 programmed into the blood gas analyser by the manufacturer. data set to filter out poor pulse quality data points, thereby to The co-oximeter was used to measure the fractional arterial obtain a good quality data set (pass-data). It was done by oxygen haemoglobin saturation of the arterial blood (SaO2). excluding any triplicate data sets: 1) with red oximeter pulse quality light reading; and 2) where the pulse quality light Statistical analysis indicated green or amber, but the SpO2 triplicate data had an SD of more than 3%. For Bland-Altman method analysis to be performed at an a ¼ To determine the performance of the pulse oximeters at 0.05 and b ¼ 0.90, this study needed a minimum of 17 paired different SaO2 ranges, the SpO2 data were compared at the data sets (i.e. SpO2 - SaO2, cSaO2 - SaO2 and SpO2 - cSaO2) to following different SaO2 or cSaO2 ranges: 0 e 100%, 70 e have power when the expected mean of differences is 3%, ex- 100% (range claimed by the manufacturer to be most accurate pected SD of differences is 3% and the maximum allowed dif- for the pulse oximeter), < 70% (range claimed by the manu- ference between methods is ±10%. facturer to be inaccurate for pulse oximetry), 70 e 79%, 80 e The SaO2 measured by the co-oximeter is accepted by the 89% and 90 e 100% for ‘all-data’ and the ‘pass-data’ sets. International Organisation for Standardisation (ISO) 80601-2- Analyses were performed using commercially available 61:2011 guidelines as the gold-standard measurement of SaO2 software MedCalc Version 19, MedCalc Software Ltd., to validate the use of pulse oximetry (DIN EN, 2011). The cSaO2 (Belgium) and GraphPad Prism, Version 7, (GraphPad Soft- calculated by the blood gas analyser is used as a ‘clinical stan- ware, CA, USA). dard’ (Whitehair et al. 1990; Reiners et al. 2018). Both mea- sures (SaO2 and cSaO2) were used to assess the accuracy and Results utility of the pulse oximeters in this study. The Bland-Altman method for multiple observations (Bland & Altman 2007) was Quality of data collected used to determine the accuracy and precision of the measure- ments recorded using the three monitors. More specifically: 1) All immobilization procedures were successful (Pfitzer et al. how SpO2 measured by the pulse oximeters at each site 2019). Overall, there were 167 paired SaO2 readings compared with the SaO2; 2) how cSaO2 calculated by the blood measured by the co-oximeter and the SaO2 data ranged from gas analyser compared with the measured SaO2; and 3) how 16.3% to 99.3%. There were 194 paired cSaO2 readings SpO2 measurements at each site compared with the cSaO2. calculated by the blood gas analyser and the data ranged from The Bland-Altman method calculates the bias [the measure 15.5% to 99.9%. The SaO2 readings had a good agreement of accuracydby calculating the mean difference between the (bias 1, LOA 5 to 7) with cSaO2 readings (Fig. 2). two variables (methods) that are compared, i.e. SpO2 - SaO2, All probe sites gave good pulse quality readings, i.e. per- cSaO2 - SaO2 and SpO2 - cSaO2], the precision (a measure of centage readings that had a good pulse qualitydgreen light, 326 © 2020 Association of Veterinary Anaesthetists and American College of Veterinary Anesthesia and Analgesia. Published by Elsevier Ltd. All rights reserved., 47, 323e333
Reliability of pulse oximetry in impala TK Mtetwa et al. (Tables 1 and 2), with the highest values obtained from the tail readings, however, the Arms indicated that pulse oximetry was and the lowest from the ear. The ear, the cheek and the rectum accurate for ‘all-data’ and ‘pass-data’. Below these ranges, had SpO2 readings indicating poor pulse quality (poor signal pulse oximetry was inaccurate and imprecise. In the manu- strengthdred light). facturer’s claimed performance range (70e100%) the read- Although the tail gave the best quality SpO2 readings, there ings were accurate but imprecise (Table 4). was a larger variability in the readings obtained from the tail, and likewise from the ear (triplicate readings with SD>3%, Tables 1 Ear and 2) compared with the other two sites (rectum and cheek). When compared with SaO2 (Tables 3 and 4), ‘all-data’ readings were accurate but imprecise in all ranges except Pulse oximeter performance at each site within the 70e79% and 90e100% ranges, where the The mean difference between the two variables (methods) that readings were both inaccurate and imprecise. These findings were compared, i.e. SpO2eSaO2, cSaO2eSaO2 and SpO2ec- were similar to the ‘pass-data’. At the range of 70e100%, the SaO2, were normally distributed. pulse oximeter gave accurate but imprecise readings for ‘all- data’ and ‘pass-data’ (Fig. 3). When compared with cSaO2, the Tail ear probe was inaccurate and imprecise at all ranges. Compared with SaO2 (Tables 3), the pulse oximeter had an Rectum acceptable performance (i.e. it was accurate and precise) in the SaO2 range of 90e100% for ‘all-data’. In the SaO2 range of Within the SaO2 ranges of 80e89% and 90e100%, the pulse 80e89%, below this range (70e79% and
Reliability of pulse oximetry in impala TK Mtetwa et al. a b 40 Mean difference [SpO2-SaO2 (% )] 20 0 –20 –40 c d 40 Mean difference [SpO2-cSaO2 (% )] 20 0 –20 –40 Tail Rectum Ear Cheek Tail Rectum Ear Cheek Figure 3 The performance of the peripheral oxygen haemoglobin saturation (SpO2; measured by pulse oximetry) compared with arterial oxygen haemoglobin saturation (SaO2; measured by the co-oximeter) (graphs a and b) and arterial oxygen haemoglobin saturation (cSaO2; calculated by blood gas analyser) (graphs c and d) in the range of 70 e 100% (manufacturer’s claimed performance range of the pulse oximeters) shown on the y axis, at four probe attachment sites in 16 impala, shown on the x axis. a and c indicate ‘all-data’ sets (i.e. when exclusion criteria were not applied) and b and d indicate “pass-data” sets (i.e. when exclusion criteria were applied). The data were analysed using the Bland-Altman method for multiple comparisons per individual. The mean differences (biaseblack squares) (SpO2-SaO2; SpO2-cSaO2) and the limits of agreement [mean ± 1.96 standard deviation (SD) indicated by the error bars] are plotted for the four probe attachment sites. The black line at 0 indicates a bias of 0; the closer the mean differences are to this line, the greater the accuracy of the pulse oximeter measurements and the smaller the limits of agreement, the greater the precision of these measurements. Cheek provided good pulse quality readings; however, the variability of these readings was also high. The accuracy of pulse oximetry When compared with SaO2 (Table 3), pulse oximetry was was poor when the probes were placed on the ear, inside the inaccurate and imprecise at all ranges except within the range rectum and against the cheek. The differences in the perfor- of 90e100%, where readings were inaccurate but precise for mance of the pulse oximeter at different attachment sites ‘all-data’ and ‘pass-data’. highlights the importance of appropriate probe placement When compared with cSaO2 (Table 4), the cheek probe gave when monitoring and assessing oxygen levels in free-ranging inaccurate but precise readings in the 90e100% range for wild ungulates during immobilization. ‘pass-data’. Within all the other ranges analyzed, the cheek The co-oximeter was not available on all study days, which probe gave inaccurate and imprecise readings. resulted in a smaller sample size of SaO2 measurements when Discussion compared with cSaO2 readings. However, the number of samples was still adequate for the comparisons made, and In the immobilized impala, SpO2 measurements using a com- similar findings were obtained when SpO2 values were mercial veterinary pulse oximeter device were accurate and compared with cSaO2. Many studies also compare the pulse precise indicators of oxygenation. However, it was only rate measured by the pulse oximeter with the heart rate (HR) acceptable at high SaO2 and was influenced by the probe’s of the animal to determine the quality of the data measured. If location on the impala. Pulse oximetry was most accurate and the two values are similar, this indicates that the pulse oxim- precise when the probe was placed at the base of the tail and eter is measuring a pulse and not movement artefact the saturation values were greater than 90%. The tail probe (Bohnhorst et al. 2002; Young et al. 2002). However, normal 328 © 2020 Association of Veterinary Anaesthetists and American College of Veterinary Anesthesia and Analgesia. Published by Elsevier Ltd. All rights reserved., 47, 323e333
Reliability of pulse oximetry in impala TK Mtetwa et al. Table 1 The number of time-matched paired data obtained from the co-oximeter (SaO2) and the pulse oximeters (SpO2) at the different attachment sites (ear, cheek, tail and rectum). The number of pairs for all the data (all-data) and their pulse quality are presented. Number of data pairs excluded, and the criteria of exclusion and the total number of data pairs not excluded (pass-data) are also presented. SD, standard deviation. sample size (n). Attachment site All-data n Pulse quality indicator Pass-data n (%) Exclusion criteria n (%) light colour (%) Triplicate SpO2 with SD >3* Red light (poor pulse quality)y Tail 167 Green: 98 124 (74) 43 (26) 0 (0) Amber: 2 Red: 0 Ear 134 Green: 75 91 (68) 35 (26) 8 (6) Amber: 21 Red: 4 Rectum 153 Green: 83 115 (75) 37 (24) 1 (1) Amber: 16 Red: 1 Cheek 111 Green: 88 88 (79) 22 (20) 1 (1) Amber: 11 Red: 1 * Triplicate SpO2 data excluded because the SD was more than 3% between the SpO2 triplicate data, even though the pulse quality indicated a green or amber light for the readings. y Triplicate SpO2 data excluded because one or more SpO2 readings had a poor pulse quality indicated by a red light. variability between HR measured from an electrocardiograph data was based solely on the pulse quality determined by the and the pulse rate measured by pulse oximetry have been internal light emitting diode function and the variability of the observed in cats, dogs and horses (Mathews et al. 2003). data. In some instances, more than 30% of good pulse quality Therefore, using HR alone to assess data quality can have its data were excluded because they were highly variable, possibly limitations. Most studies that validated their readings using indicating an error in the measurement of pulse quality, or pulse rate did not take pulse quality, or its variability, into acute changes in SaO2 of the animal at the time of measure- consideration, and this practice may limit the ability to make ment or movement. valid comparisons. We were unable to record the HR at the Previously, the accuracy of pulse oximetry has been assessed same time as our data collection. Therefore, the quality of our by either comparison of SpO2 values with calculated SaO2 Table 2 The number of time-matched paired data obtained from the blood gas analyser (cSaO2) and the pulse oximeters (SpO2) at the different attachment sites (ear, cheek, tail and rectum). The number of pairs for all the data (all-data) and their pulse quality are presented. Number of data pairs excluded, and the criteria of exclusion and the total number of data pairs not excluded (pass-data) are also presented. SD, standard deviation. sample size (n). Attachment site All-data n Pulse quality indicator Pass-data n (%) Exclusion criteria n (%) light colour (%) Triplicate SpO2 with SD >3* Red light (poor pulse quality)y Tail 194 Green: 97 134 (69) 60 (31) 0 (0) Amber: 3 Red: 0 Ear 139 Green: 72 80 (58) 50 (36) 9 (6) Amber: 24 Red: 4 Rectum 176 Green: 80 129 (73) 46 (26) 1 (1) Amber: 19 Red: 1 Cheek 136 Green: 87 105 (77) 30 (22) 1 (1) Amber: 12 Red: 1 * Triplicate SpO2 data excluded because the SD was more than 3% between the SpO2 triplicate data, even though the pulse quality indicated a green or amber light for the readings. y Triplicate SpO2 data excluded because one or more SpO2 readings had a poor pulse quality indicated by a red light. © 2020 Association of Veterinary Anaesthetists and American College of Veterinary Anesthesia and Analgesia. Published by Elsevier Ltd. All rights 329 reserved., 47, 323e333
Reliability of pulse oximetry in impala TK Mtetwa et al. Table 3 The performance of the pulse oximeter readings (SpO2) compared with the gold standard co-oximetry (SaO2) at different attachment sites in immobilized impala (number of impala ¼ 16). SaO2, oxygen haemoglobin saturation measured by co-oximetry. Attachment site SaO2 Ranges ‘All-data’ ‘Pass-data’ n Bias Precision Arms LOA n Bias Precision Arms LOA Tail 0e100% 167 6 7 10 21, 8 124 6 7 9 19, 7 70e100% 144 4 5 6 14, 5 110 4 5 6 13, 5
Reliability of pulse oximetry in impala TK Mtetwa et al. Table 4 The performance of the pulse oximeter readings compared with the blood gas analyser (cSaO2) at different attachment sites in immobilized impala (number of impala ¼ 16). Arms, area root mean squares; LOA, limits of agreement (maximum and minimum); cSaO2, oxygen haemoglobin saturation was calculated from arterial partial pressure of oxygen as measured by EPOC blood gas analyser. Attachment site cSaO2 Ranges (%) ‘All-data’ ‘Pass-data’ n Bias (%) Precision (%) Arms LOA n Bias (%) Precision (%) Arms LOA Tail 0e100 194 6 9 11 24, 12 134 5 8 9 20, 10 70e100 157 3* 6 7 16, 9 113 3* 5 6 13, 7
Reliability of pulse oximetry in impala TK Mtetwa et al. placement, Barton et al. (1996) used a sensor specifically Authors’ contributions designed for the rectum and secured it using a soft silicone TM: preparation of equipment, data collection, statistical plug, factors that may account for their consistent results. analysis and preparation of manuscript. LM, GZ: idea of the However, a similar study performed by Giguere et al. (2014) study, data collection, data interpretation and preparation of also indicated larger bias (23%) at saturations below 85%, manuscript. LL: data collection and edited manuscript. SP: data when the probe was placed inside the rectal mucosa of collection and edited manuscript. anaesthetized neonatal foals. Etorphine is known to cause severe hypoxaemia in immo- Conflict of interest statement bilized impala (Meyer et al. 2010; Zeiler et al. 2015). During immobilization, overestimation of pulse oximeter measure- SP and LL work for Wildlife Pharmaceuticals, the company ments is arguably worse than underestimation, since it may that sells the drugs that were used to immobilize the animals in give the impression that an animal has normal oxygen levels this study. The aim of this study was not to determine the ef- when it is hypoxaemic in reality. In this study, probes that fects of these drugs and the authors do not advocate their were placed in the rectum and on the cheek overestimated specific use in this manuscript. The other authors declare no SaO2, therefore these sites should be avoided. This fact is conflict of interest. particularly important especially in clinically ill animals. References Conclusion Barton LJ, Devey JJ, Gorski S et al. (1996) Evaluation of trans- We have shown that pulse oximetry, using a commercial mittance and reflectance pulse oximetry in a canine model of veterinary device, is valid in immobilized impala, but only hypotension and desaturation. J Vet Emerg Crit Care 6, 21e28. when oxygen haemoglobin saturations are high (>90%) and Bardell D, West W, Senior JM (2017) Evaluation of a new handheld when a transflectance probe is used at the base of the tail. Since point-of-care blood gas analyser using 100 equine blood samples. chemical immobilization often results in severe hypoxaemia in Anaesth Analg 44, 1e9. impala, the interpretation of pulse oximetry readings to mea- Batchelder PB, Raley DM (2007) Maximizing the laboratory setting sure haemoglobin oxygen saturation should be done with for testing devices and understanding statistical output in pulse caution. The authors recommend the use of additional moni- oximetry. Anaesth Analg 105, S85eS94. Bland JM, Altman DG (1986) Statistical methods for assessing toring equipment such as blood gas analysis or co-oximetry. agreement between two methods of clinical measurements. However, given that pulse oximetry is cost-effective, nonin- Lancet 1, 307e310. vasive and rapidly measures trends in blood oxygenation, the Bland JM, Altman DG (2007) Agreement between methods of development of more reliable pulse oximeters for use in wild measurement with multiple observations per individual. antelope species is warranted. Using the tail base as an optimal J Biopharm Stat 17, 571e582. site for pulse oximetry measurements in impala may help with Bohnhorst B, Peter CS, Poets CF (2002) Detection of hyperoxaemia this development and refinement of future clinical practices. in neonates: data from three new pulse oximeters. Arch Dis Child Fetal Neonatal Ed 87, F217eF219. Funding Burns PM, Driessen B, Boston R, Gunther RA (2006) Accuracy of a third (Dolphin Voyager) versus first generation pulse oximeter The study was funded by the University of Pretoria, Faculty of (Nellcor N-180) in predicting arterial oxygen saturation and pulse Veterinary Science, Department of Paraclinical Science, South rate in the anesthetized dog. Vet Anaesth Analg 33, 281e295. African Veterinary Association (Wildlife Group) and the Na- Chaffin MK, Mathews NS, Cohen ND, Carter GK (1996) Evaluation tional Research Foundation. of pulse oximetry in anaesthetised foals using multiple combi- nations of transducer type and transducer attachment site. Acknowledgements Equine Vet J 28, 437e445. Clerbaux T, Gustin P, Detry B et al. (1993) Comparative study The authors would like to thank Wildlifevets.com and of the oxyhaemoglobin dissociation curve of four mammals: Wildlife Pharmaceuticals for their generous support and help man, dog, horse and cattle. Comp Biochem Physiol 106A, with data collection. Additionally, we thank Dr Etienne Bason 687e694. and Penelope Miya for their help. We also thank Kyron Coghe J, Uystepruyst C, Bureau F, Lekeux P (1999) Non-invasive Laboratories and Ascendis Medical for loaning the Nonin assessment of arterial haemoglobin oxygen saturation in cattle pulse oximeters and Co-oximeter devices, and the University by pulse oximetry. Vet Rec 145, 666e669. of Pretoria, Faculty of Veterinary Science, Department of Dawson J, Bastrenta P, Cavigioli F et al. (2014) The precision and Paraclinical Science and National Research Foundation for accuracy of Nellcor and Masimo oximeters at low oxygen satu- rations (70%) in newborn lambs. Arch Dis Child Fetal Neonatal funding. Ed 99, F278eF281. 332 © 2020 Association of Veterinary Anaesthetists and American College of Veterinary Anesthesia and Analgesia. Published by Elsevier Ltd. All rights reserved., 47, 323e333
Reliability of pulse oximetry in impala TK Mtetwa et al. DeMeulenaere S (2007) Pulse oximetry: uses and limitations. Reiners JK, Rossdeutsher W, Hopster K et al. (2018) Development J Nurs Pract 3, 312e317. and clinical evaluation of a new sensor design for buccal pulse Fahlman A (2014) Oxygen therapy. In: Zoo Animal and Wildlife oximetry in horses. Equine Vet J 50, 228e234. Immobilization and Anesthesia (2nd edn). West G, Heard D, Serfontein GL, Jaroszewicz AM (1978) Estimation of gestational age Caulkett N (eds). John Wiley & Sons, Inc., NJ, USA. pp. 69e81. at birth: comparison of two methods. Arch Dis Child 53, 59e511. Grap MJ (2002) Pulse oximetry. Critical Care Nurse 22, 69e74. Uystepruyst CH, Coghe J, Bureau F, Lekeux P (2000) Evaluation of Giguere S, Sanchez LC, Shih A (2014) Accuracy of calculated arterial accuracy of pulse oximetry in newborn calves. Vet J 159, 71e76. saturation in oxygen in neonatal foals and effect of monitor, sensor, Quinn CT, Raisis AL, Musk GC (2013) Evaluation of Masimo signal site of sensor placement, and degree of hypoxemia on the accuracy extraction technology pulse oximetry in anaesthetized pregnant of pulse oximetry. J Vet Emerg Crit Care 25, 539. sheep. Vet Anaesth Analg 40, 149e156. Grubb TL, Anderson DE (2017) Assessment of clinical application of van de Louw A, Cracco C, Cerf C et al. (2001) Accuracy of pulse pulse oximetry probes in llamas and alpacas. Vet Med Sci 3, oximetry in the intensive care unit. Intens Care Med 27, 169e175. 1606e1613. Haymerle A, Knauer F, Walzer C (2016) Two methods to adapt the Wong DM, Alcott CJ, Wang C et al. (2011) Agreement between human haemoglobin-oxygen dissociation algorithm to the blood arterial partial pressure of carbon dioxide and saturation of he- of the white rhinoceros (Ceratotherium simum) and to determine moglobin with oxygen values obtained by direct arterial blood the accuracy of pulse oximetry. Vet Anaesth Analg 43, 566e570. measurements versus noninvasive methods in conscious healthy ISO 80601-2-61:2011 DIN EN (2011) Medical electrical equip- and ill foals. J Am Vet Med Assoc 239, 1341e1347. ment- Part 2-61: particular requirements for basic safety and Whitehair KJ, Watney GCG, Leith DE, Debowes RM (1990) Pulse essential performance of pulse oximeter equipment (ISO 80601- oximetry in horses. Vet Surg 19, 243e248. 2-61:2011). German version EN ISO 80601-2-61:2011. Young SS, Skeans SS, Lamca JE, Chapman RW (2002) Agreement Lee WW, Mayberry K, Crapo R, Jensen RL (2000) The accuracy of of SpO2, SaO2 and ScO2 in anesthetized cynomolgus monkeys pulse oximetry in the emergency care unit. Am J Emerg Med 18, (Macaca fascicularis). Vet Anaesth Analg 29, 150e155. 427e431. Zeiler GE, Meyer LCR (2017a) Comparison of thiafentanil- Martin-Jurado O, Bektas R, Fahrton A et al. (2011) Comparison of medetomidine to etorphine-medetomidine immobilization of im- the effects of racemic ketamine and s-ketamine for anesthesia in palas (Aepyceros melampus). J S Afr Vet Assoc 88, 1e8. rheem gazelles (Gazella subgutturosa marica) and subgutturosa Zeiler GE, Meyer LCR (2017b) Captive management of wild impala gazelles (Gazella subgutturosa subgutturosa). Am J Vet Res 72, (Aepyceros melampus) during intensive immobilization and 1164e1170. anaesthesia study trials. J Zoo Wildl Med 48, 1058e1071. Mathews NS, Hartke S, Allen JC (2003) An evaluation of pulse Zeiler GE, Stegmann GF, Fosgate G et al. (2015) Etorphi- oximeters in dogs, cats and horses. Vet Anaesth Analg 30, 3e14. neeketamineemedetomidine total intravenous anaesthesia in Meyer LCR, Hetem RS, Fick LG et al. (2008) Thermal, cardiore- wild impala (Aepyceros Melampus) of 120-minute duration. Vet spiratory and cortisol responses of impala (Aepyceros melampus) to Anaesth Analg 6, 755e766. chemical immobilization with 4 different drug combinations. J S Afr Vet Assoc 79, 121e129. Received 18 February 2019; accepted 6 August 2019. Meyer LCR, Hetem RS, Fick LG et al. (2010) Effects of serotonin Available online 4 February 2020 agonists and doxapram on respiratory depression and hypoxemia in etorphine-immobilized impala (Aepyceros Melampus). J Wildl Dis 46, 514e524. Oldham HG, Bevan MM, McDermott M (1979) Comparison of new miniature wright peak flow meter with the standard wright peak Appendix A flow meter. Thorax 34, 807e808. Pedersen T, Hovhannisyan K, Moller AN (2009) Pulse oximetry for perioperative monitoring. Cochrane Database Syst Rev 4, qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi X ffi CD002013. Arms ¼ ni¼n ðSpO2 SaO2 Þ2 n (1) Pfitzer S, Laubscher L, Meyer L et al. (2019) Dose-effect study of the serotonin agonist R-8-OH-DPAT on opioid-induced respiratory qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Xn depression in blesbok (Damaliscus pygargus philipsi) and impala Arms ¼ i¼n ðSpO2 SaO2 Þ2 n (Aepyceros melampus). Vet Anaesth Analg 46, 796e806. rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi X . Proulx J (1999) Respiratory monitoring: Arterial blood gas analysis, pulse oximetry, and end-tidal carbon dioxide analysis. Clin Tech Arms ¼ ði ¼ nÞ^nðSpO2 SaO2 Þ^2 n Small Anim Prac 14, 227e230. sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Pn 2 i¼n ðSpO2 SaO2Þ Arms ¼ n © 2020 Association of Veterinary Anaesthetists and American College of Veterinary Anesthesia and Analgesia. Published by Elsevier Ltd. All rights 333 reserved., 47, 323e333
You can also read